fix inference.py to be compatible with new Dataset class

src/inference.py

@@ -8,15 +8,23 @@
 from dataset_class import IntracranialDataset
 from torch.utils.data import DataLoader
 from tqdm import tqdm
+import os
 
 # Point to the relevent test label data and DICOM files
-testing_data = pd.read_csv('../input/label_dataset/prototype_test_labels.csv') # CSV with the ground-truth labels
-data_path = '../input/data' # folder containing dicom images
+data_path = '../holdout_test_data' # folder containing dicom images
+if os.path.exists('../inf_output'):
+    print('inference output path exists')
+else:
+    print("\noutput directory does not exist")
+    os.mkdir('../inf_output/')
+    print("Directory ../inf_output/ created\n")
 
 #Inference
 device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
 # initialize model
 model_inf = models.model_fxn(pretrained = False, requires_grad = False).to(device)
+model_inf = torch.nn.DataParallel(model_inf).to(device)
+
 # load model checkpoint
 checkpoint = torch.load('../output/model.pt')
 # load model weights state_dict
@@ -26,15 +34,13 @@
 
 
 #
-label_options = testing_data.columns.values[1:-1]
+label_options = ['any','epidural','intraparenchymal','intraventricular','subarachnoid','subdural']
 
-test_data = IntracranialDataset(
-    testing_data, path = data_path, train = False, test = True
-)
+test_data = IntracranialDataset(path=data_path, train=False, test=True)
 
 test_loader = DataLoader(
     test_data,
-    batch_size = 1,
+    batch_size = 512,
     shuffle=False
 )
 
@@ -67,10 +73,10 @@
 ), id_vars = ['Image'], var_name = 'subtype', value_name = 'ground_truth')
 
 # Save a .csv file with the inference results (label predictions) and ground-truth labels
-pd.merge(predictions_df, ground_truth_df).to_csv('../output/inference_results.csv', index = False)
+pd.merge(predictions_df, ground_truth_df).to_csv('../inf_output/inference_results.csv', index = False)
 
 # Format and save .csv with ID and predicted labels for Kaggle submission
-kaggle_submission = pd.merge(predictions_df, ground_truth_df)
-kaggle_submission['ID'] = kaggle_submission['Image'] + '_' + kaggle_submission['subtype']
-kaggle_submission = kaggle_submission[['ID', 'pred_label']].rename(columns = {'pred_label': 'Label'})
-kaggle_submission.to_csv('../output/kaggle_submission.csv', index = False)
+#kaggle_submission = pd.merge(predictions_df, ground_truth_df)
+#kaggle_submission['ID'] = kaggle_submission['Image'] + '_' + kaggle_submission['subtype']
+#kaggle_submission = kaggle_submission[['ID', 'pred_label']].rename(columns = {'pred_label': 'Label'})
+#kaggle_submission.to_csv('../output/kaggle_submission.csv', index = False)